A deep learning approach to focal cortical dysplasia segmentation in children with medically intractable epilepsy
Bibliographic record
Abstract
Paediatric epilepsy is one of the most common neurological disorders and has major impact on the cognition and quality of life of children. Focal Cortical Dysplasia (FCD) is one of the most common causes of medically intractable epilepsy. FCD may be amenable to surgical resection to achieve seizure freedom. By improving the detection of lesions such as FCD, the surgical outcome of these patients can be improved. The MRI features of FCD can be subtle and may not be detected by visual inspection. Patients with epilepsy who have normal Magnetic Resonance Imaging (MRI) are considered to have MR-negative epilepsy. Recent advances in deep learning techniques hold the potential to improve the detection of FCD lesions. The advantage of deep learning techniques, specifically Convolutional Neural Networks (CNN), and Fully Convolutional Networks (FCN) are that they are built to extract detailed features in images with minimal user involvement. Therefore, we set to develop a model, which takes an MRI, classifies whether it is FCD or not and outputs the lesion???s location in FCD cases. Also, another potential method is by considering information from different MRI sequences such as T1-weighted, T2-weighted and FLAIR simultaneously, since the MRI features of FCD may be more apparent on one sequence but not another. There are several challenges associated with training a model, such as lack of ground-truth, and unbiased data. We will address the ground-truth issue by building a pixel-level ground truth, and the unbiased data problem by sampling the healthy data to match the number of lesional data. We developed 5 models working on different inputs and generating coarse to fine localization of the lesion and compared their performances on MR-positive and MR-negative subjects. Our data was acquired from the SickKids hospital in Toronto and consisted of 56 MR-positive, 24 MR-negative, and 15 healthy patients. Our multi-sequence model successfully classified all healthy cases. Furthermore, it detected 55 MR-positive and 22 MR-negative subjects. We obtained 74% and 68% lesion coverage for MR-positive and MR-negative subjects, respectively. Based on our experiments FCN is a promising tool in segmentation and detection of FCD cases given the MRI data.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".